Download Nicholas Crafts – The threat of secular stagnation in Europe

Nicholas Crafts
Professor
University of Warwick
The threat of secular stagnation in Europe:
an historical perspective
Long-term secular stagnation is usually seen as an issue of slow trend growth but, in the tradition of Alvin Hansen, it can be regarded as primarily a problem of persistent high unemployment. Trend productivity growth appears to have slowed markedly in Europe recently but this
may not be a guide to the future, if robot technology comes through strongly. In that case,
however, unemployment of low-skill workers may be a serious secular-stagnation challenge. An
adequate response to secular stagnation through fiscal stimulus is infeasible but improved
supply-side policies in product and labour markets could provide an answer.
JEL classifications: J68, N14, O52
Keywords: robot, secular stagnation, technological progress, trend growth, unemployment
1 Introduction
The concept of secular stagnation, which
dates back to the 1930s, has been revived and was recently the topic of a recent e-book (Teulings and Baldwin,
2014). The idea of secular stagnation as
put forward by Summers (2014) is one
of a tendency to deficient aggregate demand such that negative real interest
rates are necessary to generate enough
investment to stabilize the economy at
the NAIRU. This might be a consequence of deleveraging after the financial crisis or a savings glut. If these tendencies are persistent, the economy might
face a situation where being in a liquidity trap is the new normal (Krugman,
2014). Serious as these issues may be,
they are not the focus of this paper
which takes a longer-term perspective
as more relevant to the conference theme.
The modern concern with secular
stagnation also has a long-term version
which relates to a fear that growth prospects in Europe over the medium term
are significantly worse than anyone
would have thought before the financial
crisis. A decline in long-term trend
growth of real GDP and, especially, of
labour productivity could underlie recent weakness in the European economy and Gordon (2014) sees adverse
1
demography and an absence of new
technologies with the impact of the one
big wave of the 20th century as reasons
to be pessimistic. This could, of course,
hold down both investment demand
and the neutral real rate of interest.
In contrast, the idea of secular stagnation as understood by Alvin Hansen
and his followers in the 1930s and
1940s was a tendency to high, persistent and perhaps increasing unemploy-
ment over the long-run (Higgins,
1950). As was the case with other contemporaries, Hansen saw prolonged
stagnation as giving rise to a hard core
of long-term unemployment (1941) or
in modern jargon hysteresis effects in
the labour market.1 Hansen’s diagnosis
Observers of the British labour market in the 1930s stressed that prolonged unemployment reduced employability
through loss of skills, adverse changes in worker attitudes and unfavourable perceptions of employers (Crafts,
1987).
43 rd ECONOMICS CONFERENCE 2015
129
Nicholas Crafts
of the problem evolved over time. In
the early 1930s, in the context of a controversy over technological unemployment, he saw the problem as one of insufficient price and wage flexibility to
allow adjustment to labour-saving technological change (Hansen, 1932). In
the late 1930s, in his best-known discussion of secular stagnation, he argued
that the American economy faced a cri-
sis of under investment and deficient
aggregate demand since investment opportunities had significantly diminished with the closing of the frontier,
declining population growth and a
slowdown in technological progress
(Hansen, 1939). It was as if the United
States was faced with a lower natural
rate of growth to which the rate of
growth of the capital stock would adjust through a permanently lower rate
of investment. In the early 1950s, he
endorsed the analysis of Harrod (1948)
which saw the possibility of a dynamic
equilibrium in which the actual rate of
growth was equal to the warranted rate
but below the natural rate and unemployment was increasing as demand
failed to keep up with technological
progress (Hansen, 1951).
Thus, whereas in 1939 the fear was
technological progress was too slow
by 1951 the problem might be that it
was too fast! In any event, once the
130
persistent unemployment problem was
thought of as primarily a result of inadequate aggregate demand, a possible response was to use deficit finance to provide fiscal stimulus but then over time
secular stagnation would see a steadily
increasing stock of public debt (Samuelson, 1970). This potentially raised issues of fiscal sustainability as the public
debt to GDP ratio increased, a topic
which was explored by Domar (1944).
This paper explores the relevance of
these two long-term notions of secular
stagnation to post-crisis Europe. The
diagnosis of secular stagnation promulgated 75 years ago turned out to be
completely wrong but a similar outcome this time may be less likely. Nevertheless, there are some reasons to
believe that future trend growth can
be stronger than recent performance
seems to suggest. However, even if this
is the case, the basis may be technological progress which has a strong skill
bias and which undermines the employment prospects of low-skill workers.
This would be a serious challenge to
European labour markets and is unlikely to be amenable to a solution based
on increasing government budget deficits. Some policy implications of this
analysis are suggested.
2 Secular stagnation first time
around: Why was Alvin Hansen
wrong?
Alvin Hansen was spectacularly wrong.
The United States achieved a strong
recovery from the Great Depression
post-1933 and in the following decades
enjoyed its strongest ever growth performance. The quarter century after
World War II was a period of full
employment. Neither type of long-run
secular stagnation was experienced.
American growth was underpinned
by strong total factor productivity
(TFP) growth, both in the 1930s and
OESTERREICHISCHE NATIONALBANK
Nicholas Crafts
after World War II; TFP growth in the
private non-farm economy was 2.3%
per year in the years 1929 to 1941 and
1.9% per year during 1948 to 1973
(Field, 2011). Gordon (2000) described
these years as the crest of the big wave
in long-term productivity growth centred on advances in technologies such
as chemicals, electricity, and the internal combustion engine. Field (2011)
stressed that technological progress
was broadly based and facilitated productivity growth not just in manufacturing but transport, communications,
distribution, public utilities etc. while
the TFP growth of the 1950s and 1960s
was set in train by the national innovation system that had been established
during the interwar period. This was
based on investments in corporate
laboratories and modern universities
and delivered a significant fall in
the costs of research as experimental
science improved and the supply of
specialized human capital expanded
rapidly (Abramovitz and David, 2001).
Private investment as a share of GDP
averaged 15.6% during 1948–66 –
roughly the level of the 1929 peak –
as business responded to the opportunities created by this dynamic economy.2
Unemployment in the American
economy averaged 4.4% of the labour
force during the 1950s and 1960s, perhaps slightly below the NAIRU (Gordon, 1997), and on average only 9.9%
of the unemployed were out of work for
more than 6 months. One reason for
this was that the positive shock of
World War II reversed the adverse
hysteresis effects of the 1930s Great
Depression and the Beveridge Curve
relationship between unemployment
and vacancies in the 1950s once again
2
looked like that of the 1920s (Mathy,
2015). Another key feature of the period was the ease with which the American labour market accommodated
technological progress. If this is viewed
through the lens of a macroeconomic
production function, then a combination of rising wages, capital deepening,
and constant factor shares can be seen
as the result of labour-augmenting
technological change with an elasticity
of substitution between labour and capital less than 1 (Klump et al., 2007). In
the race between relative demand and
relative supply of college-educated
workers, from 1915 through 1980 a
very rapid increase in supply was almost matched by high-skill-augmenting
technological change which raised the
demand for these workers so that the
college wage premium decreased slowly
(Goldin and Katz, 2008). In terms of
occupations, the proportion of low skill
employment fell gradually from 40.8%
in 1950 to 36.0% in 1970 while highskill white collar jobs rose from 17.9%
to 23.4% in those same years (Katz and
Margo, 2013).
By the 1950s, the successful productivity performance of the United
States as the leading economy had by
the 1950s created a great opportunity
for rapid catch-up growth in Western
Europe which experienced a Golden
Age of growth through the early 1970s
(Crafts and Toniolo, 2008). This was
based on the rapid diffusion of American technology together with big improvements in supply-side policies including, notably, moves to greater
European economic integration stimulated initially by the conditionality of
the Marshall Plan and consolidated by
the formation of the European Economic Community.3 The productivity
In addition, demographic pessimism was confounded and ( for reasons that are not entirely understood) the baby
boom began in the late 1940s.
43 rd ECONOMICS CONFERENCE 2015
131
Nicholas Crafts
gap between Europe and the United
States was rapidly reduced.4
In this Golden Age period there
was no need for deficit-financed government expenditure to fill a deflationary gap so this Keynesian response to
secular stagnation was in abeyance. As
Matthews (1968) pointed out with
regard to the United Kingdom, a country which was not unsympathetic to
such a strategy, the strength of private
investment demand rendered such
policy actions quite superfluous.
Ironically, however, this was a period
when, under conditions of financial repression, real interest rates were very
low and generally well below the
growth rate (Allsopp and Glyn, 1999)
and significant primary budget deficits were consistent with a stable public debt to GDP ratio. For example, in
the UK case, it would have been
possible to run a primary budget deficit of 3.6% of GDP on average throughout the 1950s and 1960s (Crafts,
2015a).
3 European medium-term
growth prospects
One way to predict future mediumterm growth is to assume that recent
trend growth will continue. The trend
can be estimated using quite sophisticated time-series econometrics but the
analysis is essentially backward-looking. Since recent European growth
performance both pre- and post-crisis
has generally been disappointing, approaches of this kind are pessimistic
about future growth. This is not only
true for Europe but also to some extent
for the United States where produc3
4
5
132
tivity growth slowed down after the
information and communications
technology (ICT) boom of the late
1990s.
Two methods of trend extrapolation in current use are dynamic factor
models which use high-frequency data
to try to identify trend and cyclical
components in time series of real GDP
or real GDP per worker (Antolin-Diaz
et al., 2014) and production-function
models which infer potential growth by
estimating trends in the supply-side
sources of growth including capital and
labour inputs and TFP growth (Havik
et al., 2014). Using the former methodology, Antolin-Diaz et al. (2014) conclude that trend growth both in the
United States and also in the euro area
has gradually declined since the end
of the 20th century very largely as a
result of a fall in the trend rate of
growth of labour productivity.5 They
find that trend labour productivity
growth and labour input in the euro
area has fallen to below 1% per year
and about 0% per year, respectively,
while trend growth of real GDP in the
United States has fallen by about 1 percentage point to about 2% per year
based on roughly equal contributions
from labour inputs and labour productivity growth.
Using the production-function approach, Havik et al. (2014) also conclude that trend growth is now much
lower than pre-crisis, as is reported in
table 1. The halving of European trend
GDP growth which they report is
mainly driven by reduced labour productivity growth which in turn reflects
weaker trend TFP growth.6 The results
Badinger (2005) estimated that economic integration had raised European income levels by nearly 20% by the
mid-1970s.
Real GDP per hour worked in the EU-15 rose from 38.1% of the United States level in 1950 to 62.9% by 1973
(Crafts, 2013).
The “euro area” in this analysis is a weighted average of France, Germany and Italy.
OESTERREICHISCHE NATIONALBANK
Nicholas Crafts
for Europe are actually quite similar to
those of the dynamic factor model analysis but, while accepting a growth
slowdown, the trends inferred for the
United States are rather more optimistic with trend labour productivity
growth at 1.5% per year. This is in line
with other similar analyses (Fernald,
2014). The striking implication in table 1 is that, rather than catching up as
they did for most of the postwar period, in the “new normal” European
countries will continue to fall behind
the United States in terms of productivity levels. Moreover, Europe appears
to be at greater risk of secular stagnation than the United States.
What might a more forward-looking approach say? The best starting
point for a discussion of potential longrun trend growth for the euro area is
to ask whether the United States is
heading for secular stagnation in the
long run based on an exhaustion of
technological progress (Cowen, 2011)
with the implication that future European TFP growth, which relies heavily
on the diffusion of new American technology, will be undermined. Mainstream opinion among American economists rejects this secular stagnation
thesis. Future technological progress is
notoriously hard to predict – 1980s’
pessimism was, of course, derailed by
ICT – but even Gordon (2014), often
cited as a notorious pessimist, expects
labour productivity growth at 1.3%
per year based on TFP growth around
the average of the last 40 years. He
argues that the slowdown in technological progress has already happened and
came after the end of the one big wave
of the 2nd industrial revolution in the
early 1970s although he is sceptical of
a future acceleration and believes
6
that ICT has mostly run its course.
Notwithstanding this claim, an obvious factor underpinning American
TFP growth is likely to be continuing
progress in ICT. A careful review of
developments in ICT stresses that
semiconductor technology continues
to advance rapidly and that (qualityadjusted) prices of microprocessor
chips continue to fall steeply such that a
baseline projection is that ICT-producing sectors alone will contribute about
0.4 percentage points of TFP growth
over the next decade (Byrne et al.,
2013). Moreover, since a major result
of the ICT revolution will be the ease of
analysis of massive amounts of data,
there could be a significant acceleration
in TFP growth as R & D becomes
cheaper and more productive (Mokyr,
2014).
An alternative approach is to project future American TFP growth using
a growth model based on endogenous
innovation. If the naive models of 25
years ago were invoked, then it might
be assumed that TFP growth depended
simply on R & D expenditures a share
of GDP and since these have not fallen,
neither will future TFP growth. Unfortunately, the evidence suggests the
Growth of the capital stock (and thus the capital-deepening contribution to labour productivity growth) adjusts to
TFP growth in this model.
43 rd ECONOMICS CONFERENCE 2015
133
Nicholas Crafts
constant-returns assumption embodied
in these models is not valid (Klenow
and Rodriguez-Clare, 2005). A more
realistic approach may be the semi-endogenous growth model in Jones
(2002) in which increases in human
capital and in research intensity generate transitory rather than permanent
effects on growth. This possibly has the
quite pessimistic implication that past
TFP growth in the United States has
largely come from increases in educa-
tional attainment of the population and
expansion of the R & D sector which
cannot be expected to continue so that
future TFP growth may be much slower
(Fernald and Jones, 2014). However,
even in this model, there may be countervailing tendencies; for example,
world research intensity surely still has
the scope to rise considerably as new
nations, most obviously China, become
major players.7
On balance, this review does not
give strong support to the hypothesis
that there will be secular stagnation in
the United States based on a dramatic
decline in technological progress. This
is clearly the view of OECD (2014a), as
7
8
134
reported in table 2, which uses a catchup growth model in which growth in
the leading economy (United States)
depends on demography and technological progress while long-term TFP
growth in (follower) European countries is based on TFP growth in the
leader and a component based on
reducing the productivity gap with
the leader. The OECD projections for
European countries in table 2 are based
on the assumptions that the crisis significantly reduced the level of potential
output in the short term (Ollivaud and
Turner, 2014) but has had no adverse
effect on long-run trend growth and
gradual conditional convergence towards the leading economy depending
on institutions and policies.8 In fact,
there is also more scope for catch-up
growth in most euro area economies
than before the crisis. Real GDP per
hour worked for the euro area as a
whole as a percentage of the U.S. level
has fallen from 88.7 in 1995 to 79.9 in
2007 and 76.0 in 2013.
It is certainly possible to believe
that the OECD projections are too optimistic for two main reasons. First, it
is striking that this framework leads
OECD to expect much better TFP
growth in the euro area as a whole and
in its troubled economies compared
with pre-crisis outcomes. In particular,
this will require a much better performance in TFP growth in market services of which there is no sign as yet
(van Ark et al., 2013) and which has
been the Achilles Heel of the euro area
economies in the context of excessive
regulation and weak competition (table
3) and which has also retarded the diffusion of ICT (Cette and Lopez, 2012)
China accounted for 16.2% of world R & D in 2012 compared with 2.3% in 1996 (UNESCO Institute for
Statistics, 2014).
So the very low growth in Europe of late reflects a levels-effect adjustment resulting from the financial crisis
playing out over several years rather than lower long-term trend growth.
OESTERREICHISCHE NATIONALBANK
Nicholas Crafts
which has been notably slow in some
countries (table 3).
Second, the crisis and subsequent
weak recovery may well have weakened
the political support for market friendly
supply-side policies and strengthened
forces of populism or even extremism.
Across Europe in the 1930s, prolonged
stagnation significantly increased the
electoral prospects of right-wing extremist parties (de Bromhead et al.,
2013) which were not market-friendly.
In this context, not only might it be
reasonable to worry about recent election results but it should also be recognised that opinion polls show disappointingly low support for the market
economy in many countries which have
been hit hard by the crisis.9 It is also
well-known that the Great Depression
saw big increases in protectionism.
Eichengreen and Irwin (2010) showed
that, on average, countries which devalued had lower tariffs. They argued
that protectionism in the 1930s is best
seen as a second-best policy which was
used when the conventional macroeconomic tools, fiscal and monetary policy, were unavailable, as they are for
euro area economies today. A recent
empirical analysis confirms that weak
domestic growth and losses in competitiveness continue to be conducive to
protectionism (Georgiadis and Gräb,
2013) so it is not surprising that EU
Member States have been prominent in
imposing such measures according to
Global Trade Alert (Evenett, 2014).
This does not bode well for the implementation of the Single Market in services which is an obvious antidote to
Europe’s productivity problem in market services. Nevertheless, prima facie,
it seems that with good supply-side pol9
icies medium-term growth prospects
in the euro area are better than the secular stagnation scenario might seem to
suggest.
4 Technological progress and
unemployment: Is this the real
secular stagnation threat?
The major concern of the original writers on secular stagnation was a future
of high and persistent unemployment.
This has not been the focal point of current debate but it deserves to be taken
seriously. The long hiatus in economic
growth in the euro area during the crisis and its aftermath may have significant hysteresis effects and the impact of
technological progress may be less benign than was the case during the early
postwar decades.
The estimates in table 4 project that
in several countries the cumulative output gap by 2016 will be over 30% of
GDP. In each of these countries a large
fraction of the unemployed are longterm. Past experience suggests that this
is a situation in which the employability
of those on the margins of the labour
force declines and, as a result, the
NAIRU increases. IMF (2012) estimates that an additional 1% increase in
the cumulative output gap raises the
NAIRU by 0.14 percentage points. On
this basis, table 4 reports the post-crisis
NAIRU will have risen by over 4 percentage points in Greece, Ireland, Italy,
Portugal and Spain. Moreover, the likelihood of a “positive shock” similar in
magnitude to World War II to negate
this hysteresis effect is quite small.
Since about 1980, it appears that
the implications of technological progress have become more challenging for
the labour market in OECD countries.
In response to the question “Are people better off in a free market economy?” in 2014 only 47% in Greece, 45% in
Spain and 57% in Italy agreed (Pew Research, 2014). In 2007, 67% in Spain and 73% in Italy had agreed (no
data for Greece).
43 rd ECONOMICS CONFERENCE 2015
135
Nicholas Crafts
It seems likely that in the ICT era technological progress has become capitalaugmenting and the elasticity of substitution between labour and capital has
become greater than 1 and this has reduced labour’s share of national income
by around 5% (Karabarbounis and
Neiman, 2014). Job polarization has
been a striking feature of employment
patterns in advanced economies in the
last 30 years or so with the percentages
of high-skilled (professional, managerial etc.) and low-skilled (labourers,
low-education service sector workers)
employment rising while middleskilled (clerical, blue-collar) employment has been falling. Estimates for an
aggregate of 16 European countries
show a fall of 9.27 percentage points in
the share of their middling occupations
between 1993 and 2010 against rises
for high-paying and low-paying. This
pattern is observed in most countries
with 14 of the 16 having experienced a
decline in the share of middling occupations (Goos et al., 2014). The model
estimated by these authors suggests
that this has been almost entirely due to
the factor-saving bias of technological
change rather than to offshoring with
the declining occupations being those
which entail tasks which are routine
and codifiable and thus are most amenable to computerization (Autor, 2014).
Since the early 1990s, however,
there have also been significant developments in the use of robots, a technology which raises labour productivity
substantially but also exhibits a strong
skill bias, in this case at the expense of
the low-skilled. The implication seems
to be that, thus far at least, robotics has
significantly reduced employment for
this category of worker as the substitution of workers by robots has only been
partially offset by increased demand for
output of robot-intensive production
(Graetz and Michaels, 2015). A fall in
136
real price of robots of about 80% led to
a big increase in robots per hour worked
in OECD manufacturing and added
about 0.4% per year to the growth of
real GDP per worker.
It seems very likely that the impact
of computerisation through robotics
will intensify in the near future. Frey
and Osborne (2013) estimate that 47%
of 2010 employment in the United
States has at least a 70% chance of being computerised by 2035 (table 5)
with these probabilities being strongly
negatively correlated with wages and
educational attainment of workers. Tasks
which will not be susceptible to computerization are those involving perception and manipulation, creative intelligence, or social intelligence.
If these estimates are correct, the
upside is that this technology alone
could deliver labour productivity gains
equivalent to, say, 1.5% per year over
the next 20 years. Future advances will
come in machine learning which will
be applied in mobile robotics as hitherto non-routine tasks are turned into
well-defined problems, in particular
using big data which will allow substitution of (much cheaper) robots for labour in a wide range of low-wage service occupations. It seems quite possible therefore that the issue that Europe
really confronts is actually not so much
slow technological progress but that the
skill-bias of new technologies has a big
downside in terms of a serious adjustment problem in the labour market.
If we consider the implications of
the future computerization of employment as equivalent to a an increase in
the dispersion of worker productivities,
then in an equilibrium search and
matching labour market model, the increase in equilibrium unemployment
will be greater in a setting with relatively high unemployment benefit rates
and employment protection since these
OESTERREICHISCHE NATIONALBANK
Nicholas Crafts
are labour market policies which increase the convexity of the relationship
between the unemployment rate and
skill. In a calibrated model, Mortensen
and Pissarides (1999) estimate that a
common ICT technology shock which
would raise unemployment in the
United States by about 0.4 percentage
points during 1975–1995 would have
increased unemployment by 4.8 percentage points with European Union
labour market policies.
The data reported in table 6 suggest
that many, if not all, European countries are more vulnerable to the technology shocks associated with ICT and
robotics than the United States. The
symptoms are relatively high proportions of workers with less than uppersecondary education, more generous
replacement rates, and higher levels of
employment protection. The implication is that the problem foreseen by
Hansen (1932), namely, that technological progress might create unemployment because the economy is too inflexible, may actually be a bigger threat
to Europe rather than the spectre of the
drying up of technological change proposed by Hansen (1939).
share of GDP (b) has to meet the formula b≥d(r–g) where r is the real rate
of interest on government debt and g is
the growth rate of real GDP. Second,
in the face of an increase in the public
debt ratio, the government has to be
willing to raise b (Bohn, 1998) by
enough to stabilize d. Prima facie, on
the basis of the projections in table 7,
on at least one and possibly both of
these criteria, most European countries
have some scope to use this policy approach, especially since, on these
OECD projections, real interest rates
are below growth rates for several
countries. The exceptions, unfortunately, are countries which are among
the most exposed to the hysteresis
problem such as Greece, Italy and
Portugal.
Unfortunately, there is much less
fiscal space than this since euro area
countries are committed to the fiscal
5 Fiscal sustainability and secular
stagnation
The Keynesian solution to a secularstagnation unemployment problem was
fiscal stimulus using deficit finance, fiscal sustainability permitting. Obviously, this would not be a solution to a
problem of high equilibrium unemployment resulting from skill-bias in technological progress which would require
a supply-side policy response. It might
help, however, to counteract hysteresis.
The fiscal sustainability issue can be
considered in two (related) ways. First,
in steady state to prevent an increasing
public debt to GDP ratio (d) the required primary budget surplus as a
43 rd ECONOMICS CONFERENCE 2015
compact which requires them to return
to a gross government debt ratio no
greater than 60% and to eliminate
1/20th of the excess over this level each
year. OECD (2013) calculated that to
stay within this rule for every year from
2014 to 2023, Greece will have to
maintain a primary budget surplus of
about 9% of GDP, Italy and Portugal
about 6% of GDP, and Ireland and
Spain about 3.5% of GDP and most
137
Nicholas Crafts
euro area economies will have to pursue fiscal consolidation. Moreover, if
recent trend growth is taken as a guide
to future growth rates (table 1), the required primary budget surpluses will
be considerably higher and fiscal sustainability may come into question in
some countries as the political feasibility of running adequate primary budget
surpluses becomes doubtful (Buiter and
Rahbari, 2013). Rather than fiscal stimulus being a potential antidote to secular stagnation it seems more likely that
secular stagnation will undermine the
euro area’s fiscal rules.
6 Policy implications
Although Hansen and his followers of
70 years ago thought in terms of stimulating aggregate demand through
Keynesian deficit finance, the appropriate response to long-term secular stagnation of whatever type is to improve
supply-side policies.
If secular stagnation in the guise of
slow growth is a danger, long-run
growth prospects can be improved by
pro-market reforms that raise future
TFP growth and investment as happened through European economic integration from the 1950s through the
10
138
1990s (Crafts, 2015b). It is possible to
emulate the success of these decades
through completion of the Single Market in particular with regard to services
where barriers remain high and have
not been significantly reduced in recent
years (Fournier, 2014). Estimates from
a dynamic general equilibrium suggest
that the impact could be considerable
adding perhaps 1% to the growth rate
of large euro area economies over ten
years (Aussilloux et al., 2011).10
A key focal point of policies to improve productivity growth should be to
facilitate the diffusion of technology
from the frontier, as the experience of
relatively slow adoption of ICT in some
European countries in the pre-crisis
period underlines. Recent research into
the ability of follower countries to capitalize on innovations made by the leader
suggests that investments in knowledge-based capital (both managerial
and R & D), innovation policies that
enhance the absorptive capacity of
firms, and a policy framework that supports the efficient reallocation of resources in response to new opportunities are all important in underpinning
diffusion (Saia et al., 2015).
In this context, it is important to
note that the process of creative destruction clearly works much less well
in many European countries than in the
United States as is witnessed by processes of entry and exit of firms and the
much stronger growth rate of successful American start-ups (Encaoua,
2009). A corollary of this is that, on average, countries in the European
Union, especially in Southern Europe,
are much inferior to the United States
in shifting employment away from less
productive towards more productive
firms and this may account for as much
These are, in fact, likely to be significant underestimates of the possible gains because the model does not capture
the productivity implications of greater competition.
OESTERREICHISCHE NATIONALBANK
Nicholas Crafts
as 20 percentage points of the labour
productivity gap between the EU and
the USA. Barriers to entry and strict
employment protection legislation disproportionately reduce the efficiency of
labour allocation in high turnover and
more innovative sectors (Andrews and
Cingano, 2014).
In this context, it would be most
unfortunate if in the face of lobbying
by vested interests, policymakers’ response to new technologies is to try to
slow down their diffusion, as the example of Uber might lead us to fear.
Rather, if secular stagnation in terms of
persistent high unemployment is identified as a major threat, then labour-market reforms will have a central role especially in terms of mitigating the impact of skill-bias in new technologies.
The prospect of substantial displacement of low wage labour in the service
sector creates a new challenge which is
likely to require well-designed active
labour market policies together with
stricter unemployment benefit rules
(Martin, 2014). Given the prospect of
a major disruptive new technology, it
is important that regulations which
impede the reallocation of labour are
not strengthened (Haltiwanger et al.,
2014).
7 Conclusions
Long-term secular stagnation is generally interpreted these days as very weak
trend growth but, in the spirit of Alvin
Hansen and his followers, it might better be conceptualised as a problem of
high and persistent unemployment.
Trend productivity growth appears to
have fallen significantly since the turn
of the century but future technological
change might easily surprise on the
43 rd ECONOMICS CONFERENCE 2015
upside. If this does transpire, however,
it could imply a serious risk of higher
unemployment as computerisation leads
to job losses in low skilled occupations.
While the Keynesians who pioneered the idea of secular stagnation
saw demand – side policies based on fiscal stimulus as the policy response of
choice, the right call is to improve supply-side policies in both labour and
product markets. Indeed, at present it
seems more likely that secular stagnation in terms of slow long-term growth
will undermine the euro area’s fiscal
compact than that fiscal stimulus is a
plausible solution to secular stagnation.
Key priorities in supply-side policy
include moves to improve labour market flexibility and reduce the vulnerability of Europe to skill-bias in technological progress. This will entail improving the skills of the labour force
and also reducing employment protection and unemployment benefits. It
would be a Pyrrhic victory to “solve”
this potential labour market problem
by obstructing the adoption of new
technology. To address problems of
slow productivity growth a key focal
point is to facilitate the diffusion of new
technology in particular by increasing
investments in knowledge-based capital
and by reducing obstacles to creative
destruction.
It is far too soon to tell whether secular stagnation is the future for Europe
but the risk is surely higher than in the
1930s and 1940s. It does seem clear,
however, that European countries generally are much more exposed to risks
of secular stagnation than is the United
States even though it is the Americans
who raised the alarm.
139
Nicholas Crafts
References
Abramovitz, M. and P. A. David. 2001. Two Centuries of American Macroeconomic Growth:
from Exploitation of Resource Abundance to Knowledge-Driven Development. Stanford
Institute for Economic Policy Research Discussion Paper No. 01-05.
Allsopp, C. and A. Glyn. 1999. The Assessment: Real Interest Rates. In: Oxford Review of
Economic Policy 15(2). 1–16.
Andrews, D. and F. Cingano. 2014. Public Policy and Resource Allocation: Evidence from Firms
in OECD Countries. Economic Policy 78. 253–296.
Antolin-Diaz, J., T. Drechsel and I. Petrella. 2014. Following the Trend: Tracking GDP when
Long-Run Growth is Uncertain. CEPR Discussion Paper No. 10272.
Autor, D. 2014. Polanyi’s Paradox and the Shape of Employment Growth. NBER Working Paper
No. 20485.
Aussilloux, V., H. Boumellassa, C. Emlinger and L. Fontagne 2011. The Economic Consequences for the UK and the EU of Completing the Single Market. BIS Economics Paper No.
11.
Badinger, H. 2005. Growth Effects of Economic Integration: Evidence from the EU Member
States. In: Review of World Economics 141. 50–78.
Bohn, H. 1998. The Behavior of U.S. Public Debt and Deficits. In: Quarterly Journal of Economics.
113. 949–963.
Buiter, W. and E. Rahbari. 2013. Why Do Governments Default and Why Don’t They Default
More Often? CEPR Discussion Paper No. 9492.
Byrne, D., S. Oliner and D. Sichel. 2013. Is the Information Technology Revolution Over? In:
International Productivity Monitor 25. 20–36.
Cette, G. and J. Lopez. 2012. ICT Demand Behaviour: an International Comparison. In:
Economics of Innovation and New Technology 21. 397–410.
Cowen, T. 2011. The Great Stagnation. www.digamo.free.fr/cowen11.pdf
Crafts, N. 1987. Long-Term Unemployment in Britain in the 1930s. In. Economic History Review
40. 418–432.
Crafts, N. 2013. Long-Term Growth in Europe: What Difference Does the Crisis Make? In:
National Institute Economic Review 224. R14–R28.
Crafts, N. 2015a. Reducing High Public Debt Ratios: Lessons from UK Experience. In: Fiscal
Studies Forthcoming.
Crafts, N. 2015b. West European Economic Integration since 1950: Implications for Trade and
Income. University of Warwick CAGE Working Paper No. 219.
Crafts, N. and G. Toniolo. 2008. European Economic Growth 1950–2005: An Overview.
CEPR Discussion Paper No. 6863.
De Bromhead, A., B. Eichengreen and K. H. O’Rourke. 2013. Political Extremism in the
1920s and 1930s: Do German Lessons Generalize? In: Journal of Economic History 73. 371–406.
Domar, E. D. 1944. The ‘Burden of the Debt’ and the National Income. In: American Economic.
Review 34. 798–827.
Eichengreen, B. and D. Irwin. 2010. The Slide to Protectionism in the Great Depression: Who
Succumbed and Why? In: Journal of Economic History 70. 871–897.
Encaoua, D. 2009. Nature of the European Technology Gap: Creative Destruction or Industrial
Policy? In: Foray D.(ed.) The New Economics of Technology Policy. Cheltenham: Edward Elgar.
281–314.
Evenett, S. J. 2014. The Global Trade Disorder. London: CEPR Press.
140
OESTERREICHISCHE NATIONALBANK
Nicholas Crafts
Fernald, J. G. 2014. Productivity and Potential Output before, during and after the Great Recession. NBER Working Paper No. 20248.
Fernald, J. G. and C. I. Jones. 2014. The Future of US Economic Growth In: American
Economic Review Papers and Proceedings 104. 44–49.
Field, A. J. 2011. A Great Leap Forward: 1930s Depression and U.S. Economic Growth. New
Haven: Yale University Press.
Fournier, J-M. 2014. Reinvigorating the EU Single Market. OECD Economics Department No.
1159.
Frey, C. B. and M. A. Osborne. 2013. The Future of Employment: How Susceptible are Jobs
to Computerisation? Mimeo. Oxford Martin School.
Georgiadis, G. and J. Gräb. 2013. Growth, Real Exchange Rates and Trade Protectionism since
the Financial Crisis. ECB Central Bank Working Paper No. 1618.
Ghosh, A. R., J. I. Kim, E. G. Mendoza, J. Ostry and M. S. Qureshi. 2013. Fiscal Fatigue,
Fiscal Space and Debt Sustainability in Advanced Economies. In: Economic Journal 123. F4-F30.
Goldin, C. and L. F. Katz. 2008. The Race between Education and Technology. Cambridge.
Mass. Belknap Press.
Goos, M., A. Manning and A. Salomons. 2014. Explaining Job Polarization: Routine-Biased
Technological Change and Offshoring. In: American Economic Review 104. 2509–2526.
Gordon, R. J. 1997. The Time-Varying NAIRU and its Implications for Policy. In: Journal of Economic Perspectives 11(1). 11–32.
Gordon, R. J. 2000. Interpreting the ‘One Big Wave’ in U.S. Long-Term Productivity Growth. In:
van Ark, B., S. K. Kuipers and G. H. Kuper (eds.). Productivity, Technology and Economic
Growth. Dordrecht: Kluwer Academic Publishers. 19–65.
Gordon, R. J. 2014. The Demise of U. S. Economic Growth: Restatement, Rebuttal, and Reflections. NBER Working Paper 19895.
Graetz, G. and G. Michaels. 2015. Robots at Work. CEPR Discussion Paper No. 10477.
Haltiwanger, J., Scarpetta, S. and H. Schweiger. 2014. Cross-Country Differences in Job
Reallocation: the Role of Industry, Firm Size and Regulations. In: Labour Economics 26. 11–25.
Hansen, A. H. 1932. The Theory of Technological Progress and the Dislocation of Employment.
In: American Economic Review Papers and Proceedings 22. 25–31.
Hansen, A. H. 1939. Economic Progress and Declining Population Growth. In: American
Economic Review 29. 1–15.
Hansen, A. H. 1941. Fiscal Policy and Business Cycles. London: Routledge.
Hansen, A. H. 1951. Business Cycles and National Income. New York: Norton.
Harrod, R. F. 1948. Towards a Dynamic Economics. London: Macmillan.
Havik, K., K. McMorrow, F. Orlandi, C. Planas, R. Raciborski, W. Röger, A. Rossi, A.
Thum-Thysen and V. Vandermeulen. 2014. The Production Function Methodology for
Calculating Potential Growth Rates and Output Gaps. European Economy Economic Papers
No. 535.
Higgins, B. 1950. The Theory of Increasing Under-Employment. In: Economic Journal 60.
255–274.
IMF. 2012. Article IV Report: United Kingdom.
IMF. 2013. Historical Public Finance Database.
IMF. 2015. Fiscal Monitor.
Jones, C. I. 2002. Sources of U. S. Economic Growth in a World of Ideas. In: American Economic
Review 92. 220–239.
Karabarbounis, L. and B. Neiman. 2014. The Global Decline of the Labor Share. In: Quarterly
Journal of Economics 129. 61–103.
43 rd ECONOMICS CONFERENCE 2015
141
Nicholas Crafts
Katz, L. F. and R. A. Margo 2013. Technical Change and the Relative Demand for Skilled
Labor: the United States in Historical Perspective. NBER Working Paper No. 18752.
Klenow, P. J. and A. Rodriguez-Clare. 2005. Externalities and Growth. In: Aghion, P. and
S. N. Durlauf (eds.). Handbook of Economic Growth vol. 1A. Amsterdam: Elsevier. 817–861.
Klump, R., McAdam, P. and A. Willman. 2007. Factor Substitution and Factor-Augmenting
Technological Progress in the United States: a Normalized Supply-Side System Approach. In:
Review of Economics and Statistics 89. 183–192.
Krugman, P. 2014. Four Observations on Secular Stagnation. In: Teulings, C. and R. Baldwin
(eds.). Secular Stagnation: Facts, Causes and Cures. London: CEPR Press. 61–68.
Martin, J. P. 2014. Activation and Active Labour Market Policies in OECD Countries: Stylized
Facts and Evidence on their Effectiveness. IZA Policy Paper No. 84.
Mathy, G. P. 2015. Hysteresis and Persistent Long-Term Unemployment: Lessons from the Great.
Depression and World War 2. American University Department of Economics Working Paper.
2015-02.
Matthews, R. C. O. 1968. Why has Britain had Full Employment since the War? In: Economic
Journal 78. 555–569.
Mortensen, D. T. and C. A. Pissarides. 1999. Unemployment Responses to ‘Skill-Biased’
Technology Shocks: the Role of Labour Market Policy. In: Economic Journal 109. 242–265.
Mokyr, J. 2014. Secular Stagnation? Not in Your Life. In: C. Teulings and R. Baldwin (eds.). Secular
Stagnation: Facts, Causes and Cures. London: CEPR Press. 83–89.
OECD. 2013. Economic Outlook.
OECD. 2014a. Economic Outlook.
OECD. 2014b. Education at a Glance.
OECD. 2015. Economic Outlook.
Ollivaud, P. and D. Turner. 2014. The Effect of the Financial Crisis on OECD Potential Output.
OECD Economics Department Working Paper No. 1166.
Pew Research. 2014. Pew Global Attitudes Project. www.pewglobal.org
Saia, A., D. Andrews and S. Albrizio. 2015. Productivity Spillovers from the Global Frontier
and Public Policy: Industry Level Evidence. OECD Economics Department Working Paper 1238
Samuelson, P. A. 1970. Economics 8th Edition. New York: McGraw-Hill.
Summers, L. 2014. U.S. Economic Prospects: Secular Stagnation, Hysteresis, and the Zero Lower
Bound. In: Business Economics 49(2). 65–73.
Teulings, C. and R. Baldwin (eds.). 2014. Secular Stagnation: Facts, Causes and Cures. London: CEPR Press.
UNESCO Institute for Statistics. 2014. Research and Experimental Development Database.
www.data.uis.unesco.org.
Van Ark, B. 2011. Up the Hill and Down Again: a History of Europe’s Productivity Gap Relative
to the United States,1950–2009. In: Nordic Economic Policy Review. 27–56.
van Ark, B. Chen, V. and K. Jäger. 2013. European Productivity Growth since 2000 and
Future Prospects. In: International Productivity Monitor 25. 65–83.
142
OESTERREICHISCHE NATIONALBANK
Nicholas Crafts
Table 1
Growth of potential output
and its sources
Real
GDP
Hours
worked
Table 2
OECD future growth projections
2014–2030
GDP/hour TFP
worked
Real
GDP
% per year
Employ- GDP/ TFP
ment
worker
% per year
1995–2007
EA-12
EU-15
USA
2.0
2.2
3.0
0.6
0.6
0.8
1.4
1.6
2.2
0.8
1.0
1.4
2014–2023
EA-12
EU-15
USA
1.1
1.1
2.4
0.3
0.3
0.9
0.8
0.8
1.5
0.5
0.5
1.0
Source: Derived from Havik et al. (2014).
United States
Euro area
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Netherlands
Portugal
Spain
Sweden
UK
2.4
1.7
1.9
2.0
1.6
2.0
2.2
1.1
2.2
2.3
1.5
2.1
1.4
1.5
2.6
2.6
0.5
0.2
0.2
0.4
0.1
–0.1
0.3
–0.5
0.2
1.2
0.3
0.2
0.3
0.9
0.5
0.6
1.9
1.5
1.7
1.6
1.5
2.1
1.9
1.6
2.0
1.1
1.2
1.9
1.1
0.6
2.1
2.0
1.6
1.2
1.5
1.1
1.0
1.9
1.2
1.5
1.8
0.8
0.7
1.6
0.9
0.4
1.8
1.5
Source: OECD (2014a).
Table 3
Aspects of labour productivity growth in the market sector, 1995–2007
a) Growth accounting
b) Sectoral contributions
TFP
Labour Non ICT ICT
Quality K/HW K/HW
Y/HW
ICT
Goods Market RealY/HW
produc- produc- services location
tion
tion
% per year
Austria
Belgium
Denmark
Finland
France
Germany
Italy
Netherlands
Spain
United
Kingdom
EU-10
USA
% per year
0.1
0.2
0.1
0.1
0.3
0.0
0.1
0.4
0.4
0.0
0.4
0.1
–0.1
0.4
0.5
0.4
0.0
0.5
0.5
0.9
1.0
0.5
0.3
0.5
0.2
0.5
0.4
1.5
0.1
–0.1
2.8
0.9
0.7
–0.4
1.1
–0.6
2.2
1.7
1.0
3.3
2.0
1.7
0.4
2.1
0.6
0.4
0.2
0.3
0.4
0.4
0.3
0.8
0.5
0.9
1.0
0.6
1.2
2.6
1.6
2.6
Austria
Belgium
Denmark
Finland
France
Germany
Italy
Netherlands
Spain
United
Kingdom
EU–10
USA
0.3
0.3
0.3
1.7
0.4
0.5
0.2
0.4
0.1
1.7
0.9
0.4
1.3
0.8
0.9
0.2
0.6
0.2
0.2
0.6
0.4
0.5
0.7
0.4
0.0
1.2
0.3
–0.1
–0.1
–0.1
–0.1
0.0
0.0
–0.1
–0.2
–0.1
2.2
1.7
1.0
3.3
2.0
1.7
0.4
2.1
0.6
0.5
0.4
0.8
0.7
0.7
0.3
1.6
0.6
1.8
–0.2
–0.2
–0.2
2.6
1.6
2.6
Source: van Ark (2011).
43 rd ECONOMICS CONFERENCE 2015
143
Nicholas Crafts
Table 4
Estimates of hysteresis effect of crisis
on NAIRU
Cumulative
output gap ,
2009–2016
Predicted
change in
NAIRU
% of GDP
Percentage
points
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Netherlands
Portugal
Spain
Sweden
UK
Euro area
USA
9.4
6.0
14.9
15.5
13.1
9.5
63.7
37.7
34.4
8.7
32.1
32.9
14.2
13.4
18.7
25.1
Table 6
Exposure to skill-bias of technological
change
Low
educational
attainment
1.32
0.84
2.09
2.17
1.83
1.33
8.92
5.28
4.82
1.22
4.49
4.61
1.99
1.88
2.62
3.51
Source: Author’s calculations.
Note: Change in NAIRU estimated based on cumulative output gap
from OECD (2015) using method in IMF (2012).
Table 5
Estimates of computerisation
probabilities by 2035
% of
labour
force
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Netherlands
Portugal
Spain
Sweden
UK
USA
17
28
22
16
28
13
32
25
43
27
63
46
13
22
11
Unemployment
rate of
low educated
%
7.7
12.1
9.6
11.6
13.8
12.8
25.3
23.3
12.2
6.6
16.0
31.2
12.3
10.5
14.3
Employment
protection
Net
replacement
rate
0–6
%
2.37
1.81
2.20
2.17
2.38
2.87
2.12
1.40
2.51
2.82
3.18
2.05
2.61
1.03
0.26
72
82
87
69
68
83
46
75
78
81
78
74
67
56
51
Source: OECD (2014b), OECD Benefits and Wages database and
OECD Employment Protection database.
Notes: Low educational attainment is defined as less than upper secondary for ages 25–64 in 2012; employment protection is for
permanent workers in 2013; net replacement rate is for household with 1 earner and 2 children on 67% average wage at initial unemployment in 2013.
% of 2010 employment in USA
>0.3 but < 0.7
33
19
47
Source: Frey and Osborne (2013).
144
OESTERREICHISCHE NATIONALBANK
Nicholas Crafts
Table 7
Aspects of future fiscal sustainability
2014 d
Austria
Belgium
Denmark
Finland
France
Germany
Greece
Ireland
Italy
Netherlands
Portugal
Spain
Sweden
UK
2020 r
0.868
1.056
0.426
0.596
0.951
0.731
1.772
1.095
1.321
0.683
1.302
0.977
0.415
0.895
2030 r
1.2
1.4
1.1
1.4
2.2
1.0
6.9
3.1
3.1
1.4
5.4
4.2
2.3
3.7
2014–2030 g
1.8
1.8
1.8
2.2
2.3
1.8
3.2
1.8
2.3
1.8
2.4
2.0
2.6
3.7
1.9
2.0
1.6
2.0
2.2
1.1
2.2
2.3
1.5
2.1
1.4
1.5
2.6
2.6
Max b
Limit of d
2.0
6.3
9.2
6.5
1.0
1.6
3.9
5.4
5.3
4.2
2.4
2.9
5.0
3.5
1.873
1.684
2.087
1.845
1.761
1.758
<1.586
1.497
<1.247
1.901
<0.984
1.539
2.049
1.665
Sources:
2013 d is public debt to GDP ratio in 2013 (IMF, 2015).
2020 r and 2030 r are projected real interest rates on 10-year government bonds in 2020 and 2030, respectively (OECD, 2014a).
2014–2030 g is the projected average rate of growth of real GDP between 2014 and 2030 (OECD, 2014a).
Max b is the largest average primary budget surplus as a percentage of GDP over a 5-year period since 1980 (IMF, 2013).
Limit of d is the projected public debt to GDP ratio at which past experience indicates that the response of the primary surplus would no longer satisfy
a fiscal-sustainability criterion (Ghosh et al., 2013).
43 rd ECONOMICS CONFERENCE 2015
145